Short scales for measuring schizotypy

Abstract

Background

This study reports short scales for measuring several dimensions of schizotypy in the normal population based on a large twin sample.

Methods

The four short scales use items drawn from a longer instrument, the Oxford-Liverpool Inventory of Feelings and Experiences. Using concordance estimates from MZ and DZ pairs, the items were selected both to have a high heritability and to offer broad coverage of each trait domain.

Results

Preliminary descriptive statistics are reported for the short scales and suggest adequate reliability.

Conclusions

New scales offer a time efficient and reliable method of studying proneness to psychosis in large N designs.

Introduction

Over the past several decades many self-report scales have been developed for measuring schizotypal personality traits in non-clinical individuals (for reviews see Chapman et al., 1995, Mason et al., 1997b). Factor analytic studies of these measures have most frequently supported either three or four components, depending on the number and item content of scales included in the analyses (Claridge et al., 1996, Vollema and van den Bosch, 1995). The Oxford-Liverpool Inventory of Feelings and Experiences (O-LIFE) is an instrument based on analysis of what is probably the largest single dataset of schizotypal measures (Claridge et al., 1996). It contains four sub-scales called ‘Unusual Experiences’, ‘Cognitive Disorganisation’, ‘Introvertive Anhedonia’ and ‘Impulsive Nonconformity’. Since its introduction, the O-LIFE has enjoyed wide currency. Early work established its high internal consistency (Mason et al., 1995) and test–retest reliability (Burch et al., 1998). It has also been used in a variety of studies across many research domains, firmly establishing its construct validity as a genuine measure of schizotypal traits. Laboratory investigations have demonstrated predictable effects in relation to neuropsychological function (e.g. Avons et al., 2003); on several perceptual and attentional paradigms (e.g. Steel et al., 2002); in psychophysiological responding (e.g. Mason et al., 1997a); on reasoning tasks (e.g. Sellen et al., 2005) and in learning, notably on measures of ‘latent inhibition’ (e.g. Moran et al., 2003). Differences have also been found in hemispheric function: for language task performance (Nunn and Peters, 2001), face processing (Mason and Claridge, 1999), and handedness (Shaw et al., 2001). With clinical relevance, the O-LIFE has been used successfully to investigate schizotypy in relation to such topics as dissociative experience and childhood abuse (Startup, 1999) and paranormal beliefs and experiences as a function of mental health (Goulding, 2004). It is also important that there is evidence for a genetic basis as measures of this type are otherwise open to the criticism that they are merely phenomenologically or, in Meehl's (1993) terms, ‘pseudo-phenotypically’ related to schizophrenia. A recent large quantitative genetic analysis (Linney et al., 2003) has established convincing heritability estimates for the scales. However, at over 100 items, the full scales are arguably cumbersome and a shortened form is very desirable. We felt that it was very important to retain those aspects of schizotypy most relevant to ‘true’ proneness to disorder without losing the broad ‘flavour’ of each trait. By using a twin sample we aimed to retain, and even to maximise, the degree of genotypic variance captured by items without compromising scale content.